A

pdPseL©

K

1

dchndtrodtdmplw

z

1d

Journal of Chromatography B, 857 (2007) 9–14

Quantitative determination of tilmicosin in canine serum by highperformance liquid chromatography–tandem mass spectrometry

Michael Herrera a,1, Haiqing Ding b, Robert McClanahan b, Jane G. Owens a, Robert P. Hunter a,∗a Elanco Animal Health, A division of Eli Lilly and Co., 2001W. Main St., P.O. Box 708, Greenfield, IN 46140, USA

b Ricerca Biosciences, LLC, 7528 Auburn Road, Concord, OH 44077, USA

Received 9 May 2007; accepted 22 June 2007Available online 29 June 2007

bstract

A highly sensitive and quantitative LC/MS/MS assay for the determination of tilmicosin in serum has been developed and validated. For samplereparation, 0.2 mL of canine serum was extracted with 3 mL of methyl tert-butyl ether. The organic layer was transferred to a new vessel andried under nitrogen. The sample was then reconstituted for analysis by high performance liquid chromatography–tandem mass spectrometry. Ahenomenex Luna C8(2) analytical column was used for the chromatographic separation. The eluent was subsequently introduced to the mass

pectrometer by electrospray ionization. A single range was validated for 50–5000 ng/mL for support of toxicokinetic studies. The inter-day relativerror (inaccuracy) for the LLOQ samples ranged from −5.5% to 0.3%. The inter-day relative standard deviations (imprecision) at the respectiveLOQ levels were ≤10.1%.2007 Elsevier B.V. All rights reserved.

orpiaLrttoaamas

eywords: Tilmicosin; Canine; Liquid chromatography/mass spectrometry

. Introduction

Tilmicosin, a macrolide antimicrobial, is a semi-syntheticerivative of tylosin, and is indicated for the prevention andontrol of respiratory diseases associated with Mannheimiaaemolytica in cattle and sheep and Actinobacillus pleurop-eumoniae in swine [1]. While analytical methods have beenemonstrated for the determination of tilmicosin in recent years,he vast majority of method development has centered aroundesidue determinations in various tissues. A recent review [2]f analytical methods published in the past two decades for theetermination of macrolide antimicrobials clearly emphasizeshis point. While there were many methods that included theetermination of tilmicosin in various tissues [3–10], only fiveethods were recognized for the quantitation of tilmicosin in

lasma or serum [11–15]. Many methods that have been pub-ished for biological fluids utilize either liquid chromatographyith UV detection or microbiological analysis, obtaining vari-

∗ Corresponding author. Tel.: +1 317 655 0958.E-mail address: HUNTER ROBERT P@LILLY.COM (R.P. Hunter).

1 Present address: Eurofins/AvTech Laboratories, 6859 Quality Way, Kalama-oo, MI 49002, USA.

2

2

sU

570-0232/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.jchromb.2007.06.017

us limits of quantitation. A swine serum method [12] has beeneported to have a lower limit of quantitation (LLOQ) of 25 partser billion (ppb) using 2 mL of serum. A method for determin-ng concentrations of tilmicosin in elk serum [14], which wasn adaptation of an earlier published method [3], obtained anOQ of 10 ppb but required 5 mL of serum. Liquid chromatog-

aphy coupled with a tandem mass spectrometer (MS) was usedo determine concentrations in several tissues and confirm iden-ity of tilmicosin [6], but this method did not take full advantagef the sensitivity of LC/MS/MS as the LLOQ was establishedt the maximum residue limits (MRLs) in the various tissuesnalyzed (1 mg/kg for liver and kidney; 0.05 mg/kg for fat anduscle). Thus, the method that we report here offers significant

dvantages over previous methods in terms of sample size andensitivity to support toxicokinetic studies.

. Experimental

.1. Reagents and standards

HPLC grade methanol, acetonitrile, and water and ACS gradeodium carbonate were obtained from Fisher (Pittsburgh, PA,SA). Formic acid (96%), methyl tert-butyl ether (99+%), and

10 M. Herrera et al. / J. Chromatogr. B 857 (2007) 9–14

sin an

Ao(laIe

2

SamSCm(t

2

a

ta

ncw(s2es5sa

2

a

Fig. 1. Chemical structure of tilmico

CS grade ammonium acetate and ammonium formate werebtained from Sigma–Aldrich (St. Louis, MO, USA). Tilmicosinpurity = 90.7%) was provided by Eli Lilly and Co. (Indianapo-is, IN, USA). Erythromycin (purity = 95.6% Factor A), useds the internal standard, was obtained from MP Biomedicals,nc. (Aurora, OH, USA). The structures of both tilmicosin andrythromycin can be found in Fig. 1.

.2. Equipment

The HPLC system consisted of a LC-10ATvp pump, aCL-10Avp controller, a FCV-10ALvp switching valve, andSIL-10ADvp autosampler from Shimadzu (Duisberg, Ger-any). The HPLC was coupled to an Applied Biosystemsciex API 3000 triple-quadrupole mass spectrometer (Concord,anada) for tandem mass spectrometric detection. The chro-atographic separation was performed using a Phenomenex

Torrance, CA, USA) Luna C8(2) (30 mm × 2.0 mm, 3 �m par-icle size) analytical column.

.3. Quality control samples and standard solutions

Quality control (QC) samples were prepared just prior tossay by fortifying serum with the appropriate QC working solu-

owsm

d erythromycin (internal standard).

ions. For the canine serum the QC concentrations were preparedt 50, 150, 2500, 4500, and 5000 ng/mL.

Stock solutions for both tilmicosin and erythromycin (inter-al standard) were prepared separately in acetonitrile atoncentrations of 200 �g/mL. The tilmicosin stock solutionas subsequently diluted with [acetonitrile:water:formic acid

20:80:0.1)] to prepare working standard solutions for the canineerum assay at concentrations of 500, 1000, 2000, 5000, 10,000,0,000, and 50,000 ng/mL. To 200 �L of serum, 20 �L ofach appropriate working standard solution was added for finalerum concentrations of 50, 100, 200, 500, 1000, 2000, and000 ng/mL for the canine serum assay. The internal standardtock solution was diluted with dilution solution to 2000 ng/mLs the working solution for the assay.

.4. Sample preparation

A 200 �L aliquot of serum was fortified with 20 �L ofppropriate internal standard solution. To the sample, 40 �L

f ACN:MeOH:0.1 M NH4OAc, pH 6.3–6.4 (30:20:50 v/v/v)as added followed by 20 �L of saturated sodium carbonate

olution. The sample was then extracted with 3 mL of tert-butylethyl ether, vortexed for 5 min, and centrifuged for 5 min at

M. Herrera et al. / J. Chromatogr. B 857 (2007) 9–14 11

Table 1Accuracy and precisiona

Concentration (ng/mL) n Mean + SD (ng/mL) Relative error (%) CV (%)

Canine serum50 6 50.2 ± 3.72 0.3 7.4

64.6176

3gu

sats

2

MfAgpTatw

8tTTlf

2

pctwssc

3

3

ns

swSccst

3

tws

rcocT0

3

tsc

pd≤Li

3

4iai

2500 6 2422 ±5000 6 4702 ±

a For canine assay, data from one validation batch with six replicates.

300 × g. The ether layer was transferred by pipette to a 15 mLlass tube and evaporated to dryness at approximately 37 ◦Cnder nitrogen gas flow.

The sample was then reconstituted with 1 mL of recon-titution solution, vortexed, and centrifuged for 5 min atpproximately 3300 × g. A 100 �L sample aliquot wasransferred to HPLC autosampler vials containing 900 �L recon-titution solution and then analyzed by LC/MS/MS.

.5. Chromatographic and mass spectrometric conditions

For the chromatography, two mobile phases were employed.obile phase A was 0.5% formic acid in 3 mM ammonium

ormate water solution. Mobile phase B was 0.5% formic acid inCN. The flow rate was 0.4 mL/min. The analytes were eluted byradient LC where the initial step was composed of 15% mobilehase B and increased linearly to 50% mobile phase B at 0.5 min.he composition was switched back to the original conditionst 2.6 min and held until 5 min. The column eluent was divertedo waste before 0.9 min and after 4.1 min. An injection of 3 �Las made for each sample.To perform MS/MS analysis, m/z transitions of

69.6 → 174.1 for tilmicosin and 734.5 → 158.0 for ery-hromycin were monitored using total ion monitoring. TheurboIonSpray source was operated in positive ion mode.he following settings were used for acquisition: capil-

ary = 4500 V; source temperature = 500 ◦C; orifice = 40 V;ocusing ring = 200 V; dwell time = 300 ms.

.6. Data analysis

Peak integration of tilmicosin and the internal standard wereerformed using MacQuan version 1.7.1 software. Calibrationurves were obtained by plotting the peak area ratio of tilmicosino internal standard against the amount of tilmicosin added. Aeighted (1/x2, where x is concentration) least squares regres-

ion analysis was used to obtain a linear equation over thetandard curve range. The origin was not used in the standardurve calculations.

. Results and discussion

.1. Selectivity

Selectivity was achieved by independent separation mecha-isms of chromatography and tandem mass spectrometry. Theample was initially separated by reversed-phase HPLC and sub-

3

i

−3.1 2.7−6.0 3.8

equently introduced to the mass spectrometer where the sampleas further separated and detected by mass-to-charge ratio (m/z).ix independent sources of serum were analyzed. Representativehromatograms of matrix blank and blank with internal standardan be seen in Figs. 2 and 3, respectively, for canine serum. Noignificant interferences were observed for tilmicosin (retentionime: ∼2.2 min) or erythromycin (retention time: ∼2.3 min).

.2. Linearity

The validation batch included two sets (by duplicate injec-ion) of calibration standards. One set of the calibration standardsas run at the beginning of the analytical batch, and the other

et was run at the end of the batch.The peak area ratio of tilmicosin to the internal standard is

elated to concentration using a linear fit, with 1/x2 (where x isoncentration) weighting. The standard curves had a mean slopef 0.00133 and a corresponding intercept of −0.00065. Suffi-ient linearity was observed for curve range of 50–5000 ng/mL.he correlation coefficient (r) for the calibration curve was.9977.

.3. Accuracy and precision

The accuracy and precision results for the determination ofilmicosin in canine serum are shown in Table 1 for the canineerum assay. The validation batch included six replicates at eachoncentration of the QC samples.

The inter-day relative error (inaccuracy) for the LLOQ sam-les ranged from −5.5% to 0.3%. The inter-day relative standardeviations (imprecision) at the respective LLOQ levels were10.1%. At tilmicosin concentrations above the respectiveLOQ, the inaccuracy ranged from −7.5% to 4.3% and the

mprecision was ≤4.1% for the assay.

.4. Recovery

Extraction efficiency was evaluated at 150, 2500, and500 ng/mL of tilmicosin for the canine serum assays. Includedn Table 2 is the extraction efficiency data for the canine serumssays. Average extraction efficiencies of both tilmicosin andnternal standard were approximately 90% for this assay.

.5. Stability

Tilmicosin stability was evaluated for stock solutions andn matrix. Table 3 is a summary of the stability results gener-

12 M. Herrera et al. / J. Chromatogr. B 857 (2007) 9–14

Fig. 2. Chromatogram of extracted tilmicosin naıve (blank, no IS) canine serum. The numbers above each peak represent m/z values.

Table 2Extraction efficiency from canine serum

Concentration in serum (ng/mL) Recovery (%)

Canine serum150 89.72500 91.34500 92.8(I.S.) 10000 93.3

Table 3The stability data of tilmicosin in canine serum

Concentration (ng/mL) n Accuracy (%)

Bench-top stability (∼21 h in canine serum at ambient temperature)150 3 3.24500 3 −6.6

Freeze–thaw stability (three cycles in canine serum at −80 ◦C)150 3 0.44500 3 −2.3

Long-term storage stability (92 days in canine serum at −80 ◦C)150 3 0.24500 3 −1.4

M. Herrera et al. / J. Chromatogr. B 857 (2007) 9–14 13

nal sta

asadfd

3

ctwwda

bftct2

4

w

Fig. 3. Chromatogram of extracted tilmicosin naıve canine serum with inter

ted for tilmicosin in canine serum. Standard and working stockolution stability was demonstrated for 42 days when stored atpproximately 5 ◦C. Tilmicosin stability in canine serum wasemonstrated for approximately 21 h at ambient temperature,or three freeze–thaw cycles when frozen at −80 ◦C, and for 92ays when stored at −80 ◦C.

.6. System suitability sample preparation and test

Two separate solutions each containing 50 ng/mL of tilmi-osin and 200 ng/mL of the internal standard were preparedo evaluate system suitability. The system suitability samples

ere prepared by adding 100 �L of the 500 ng/mL tilmicosinorking solution and 100 �L of the 2000 ng/mL internal stan-ard working solution to 800 �L of reconstitution solution in anutosampler vial.

coct

ndard (erythromycin). The numbers above each peak represent m/z values.

Each of the two solutions was injected in triplicate before theatch analysis. The samples exhibited acceptable peak shapeor the analyte and internal standard. The CV% of the areas ofilmicosin and internal standard peaks was within the acceptanceriteria of less than or equal to 15%. Typical retention times forilmicosin and the internal standard were approximately 2.3 and.5 min, respectively.

. Conclusions

A validation of tilmicosin in canine serum was conductedhich provided good sensitivity, linearity, accuracy, and pre-

ision over a range of concentrations providing analysisver two orders of magnitude. The sample preparation andhromatographic separation combined with tandem mass spec-rometric detection provide sufficient selectivity, minimize the

1 romat

ea[

R[[

[

[13] A. Womble, S. Giguere, Y.V.S.N. Murthy, C. Cox, E. Obare, J. Vet. Phar-

4 M. Herrera et al. / J. Ch

ffects of matrix on the results, and use a relatively smallmount of sample compared to currently published methods11–15].

eferences

[1] R.P. Hunter, R. Hassfurther, R.F. Seneriz, J.O. Clark, J. Vet. Pharmacol.Ther. 29 (Suppl. 1) (2006) 276.

[2] M.G. Gonzalez de la Huebra, U. Vincent, J. Pharm. Biomed. Anal. 39(2005) 376.

[3] W. Chan, G.C. Gerhardt, C.D.C. Salisbury, J. AOAC Int. 77 (1994) 331.

[4] C. Leal, R. Codony, R. Compano, M. Granados, M.D. Prat, J. Chromatogr.

A 910 (2001) 285.[5] B. Delepine, D. Hurtaud-Pessel, P. Sanders, J. AOAC Int. 79 (1996) 397.[6] M. Dubois, D. Fluchard, E. Soir, Ph. Delahaut, J. Chromatogr. B 753 (2001)

189.

[

[

ogr. B 857 (2007) 9–14

[7] R. Draisci, L. Palleschi, E. Ferretti, L. Achene, A. Cecilia, J. Chromatogr.A 926 (2001) 97.

[8] R. Codony, R. Compano, M. Granados, J.A. Garcıa-Regueiro, M.D. Prat,J. Chromatogr. A 959 (2002) 131.

[9] M. Horie, H. Takegami, K. Toya, H. Nakazawa, Anal. Chim. Acta 492(2003) 187.

10] J. Wang, J. Agric. Food Chem. 52 (2004) 171.11] E.A. Abu-Basha, N.M. Idkaidek, A.F. Al-Shunnaq, Vet. Res. Commun. 31

(2007) 477.12] J. Shen, H. Jiang, S. Zhang, P. Guo, S. Ding, X. Li, Am. J. Vet. Res. 66

(2005) 1071.

macol. Ther. 29 (2006) 561.14] C. Clark, M. Woodbury, P. Dowling, S. Ross, J.O. Boison, J. Vet. Pharmacol.

Ther. 27 (2004) 385.15] J.W. Moran, J.M. Turner, M.R. Coleman, J. AOAC Int. 80 (1997) 1183.